AWS MIT Workshop

Set-Up

1. Follow the instructions given by your instructor to log into your AWS Demo Account, and to create an Amazon SageMaker Notebook Instance.
2. Select Git repositories, under Repository select "Clone a public Git repository to this notebook instance only".
3. Under Git repository URL paste https://github.com/kenstler/aws_ml_workshop.git.
4. Click Create Notebook Instance.

Once your Amazon SageMaker NoteBook Instance is ready, please click Open Jupyter.

Lab 1: AWS CLI & AMI

In this lab, we will walk through examples on how to use AWS CLI to create, list, and terminate Amazon EC2 Instances. We will also learn how to create Amazon Machine Images (AMI) that have all our installed dependencies, so we can quickly create instances from them in the future.

1. Once you're in the Jupyter Console, click on Lab1 and then AWS CLI & AMI Lab.ipynb.
2. Follow the instructions in the notebook.

Lab 2

In this lab, we will go over up to 4 SageMaker examples, covering:

• 1P algorithms on Amazon SageMaker
• Train custom models using "Bring-your-own-script" and PyTorch
• Distributed training simplified with "Bring-your-own-script", Keras, and Horovod
• Reinforcement learning for Mountain Car with Intel Coach

To access these examples, please go to your Jupyter Notebook Console and Select SageMaker Examples at the top. We will go through the following examples:

1. object_detection_recordio_format.ipynb under Introduction to Amazon Algorithms
2. pytorch_rnn.ipynb under Sagemaker Python Sdk
3. tensorflow_script_mode_horovod.ipynb under Sagemaker Python Sdk
4. rl_mountain_car_coach_gymEnv.ipynb under Reinforcement Learning